Download Document 8234789

investment differs; some studies use flows of investment, others use stocks, and still others use
production, sales, capital expenditures, exports, or employment. A third reason for
incomparability is that some studies examine mainly time series variation while others focus on
cross- country variation at a given time.
Measures of Affiliate Location
There is no definitive measure of the size of affiliate operations. The stock of investment
in a host country, the most frequently used measure because of its almost universal availability,
is a financial concept, calculated for balance-of-payments purposes and matching balance-ofpayments definitions. It does not necessarily match the location of production, employment, or
other indicators of real activity because a host country may be used as a conduit for financing
production that takes place elsewhere. For example, in the 1980s, the United States held stocks
of direct investment in the Netherlands Antilles that reached a value of negative $25 billion.
There was no corresponding production there; the negative investment was the result of a tax on
foreign borrowing that could be avoided by having the affiliates in the Netherlands Antilles
borrow abroad and relend the proceeds to their parents in the United States, turning a portfolio
capital flow into a direct investment flow that was not subject to the tax. Another example of a
mismatch between investment stock and production is that, in 1997, the United States held an
investment stock of $33 billion in Bermuda, almost as large as that in Brazil and much larger
than that in Mexico, two of the major destinations for U.S. outward FDI and locations for
overseas production by U.S. multinationals. The investment in Bermuda, as well as in several
other locations, such as the Bahamas, Panama, the Netherlands Antilles, and U.K. Islands in the
Caribbean, is almost entirely in the Finance sector, and it is doubtful whether there is any real
activity in those countries, other than incorporation, represented by these large investment
stocks. The host countries’ national income and product calculations treat these operations as
outside the national economies.
2
Several measures of the size in 1994 of U.S. affiliate operations in Asian locations are
presented in Tables 1 and 2. The data on gross product and exports refer only to majority-owned
affiliates (MOFAs) because these measures are available only for them, but we can get some idea
of the importance of the omitted 50 per cent-owned and minority-owned affiliates from data for
the other variables. Singapore, despite its small population and geographical area, attracted the
most U.S. FDI, as measured by the FDI stock, MOFA gross product, and MOFA sales, and was
also, by far, the largest source of MOFA exports. Malaysia was host to the largest U.S.
operations in terms of employment. Hong Kong was the largest as measured by total assets,
twice as large if all affiliates are counted, and Indonesia was the largest as measured by fixed
assets. The contrasts among the measures reflect mainly differences in industry composition.
The importance of labor-intensive electronic product assembly in Malaysia resulted in high
levels of employment. The large financial sector investment in Hong Kong involved a high level
of total assets but not large fixed assets. And the importance of the capital-intensive Petroleum
industry in Indonesia meant larger U.S. affiliate fixed assets there than in any other Asian
location considered here.
The size of U.S. firms’ operations should be considered with reference to the size of the
host countries in which they are located, because there is a very large range among them. Two
size measures for the locations we cover, nominal GDP and employment in 1994, are shown in
Table 3. The nominal GDP figures exaggerate the differences between poorer and richer
countries because price levels are much lower in the poorer countries than in Hong Kong or
Singapore. However, we show the nominal figures because the measures for U.S. affiliates are
also in nominal terms. In terms of employment, Singapore is the smallest of the countries, at less
than one half of one per cent of China and India. In terms of nominal GDP, however, the small
countries, Malaysia, the Philippines, and Singapore, are all well over 10 per cent of the size of
the largest, China.
3
Three measures of U.S. affiliate size relative to location size are also shown in Table 3.
One is the ratio of affiliate production (gross product) to GDP and the other two are ratios of
affiliate to total employment. The ranges in the importance of U.S. affiliates are huge. The
production ratios go from a low of about one tenth of one per cent in India and China to over 8
per cent in Singapore and the employment ratios from one hundredth of one per cent or less in
India and China to around 6 per cent in Singapore. The employment ratios are far below the
output ratios because output per worker is much higher on average in affiliates than in the host
countries as a whole.
The growth of U.S. FDI differs according to the size measures used to measure it. The
smallest growth rates are, for most countries, those measured by employment, the only measure
not affected by inflation or other price changes. The increases for nine countries, excluding
China (for which we lack 1982 data), averaged 75 per cent over the 12 years, and there were
decreases in employment in two countries, India and the Philippines (Table 4). Taking all the
growth measures, one can calculate that Korea had the highest rates of growth, on average, but
from a very low level of foreign participation, still low in 1994. U.S. affiliates also grew
relatively rapidly in Taiwan, Thailand, Singapore, and Hong Kong, and least rapidly in India,
Indonesia, and the Philippines.
There are undoubtedly many explanations for the wide differences among these countries
in the extent of U.S. FDI and in its growth over the period, and we will attempt a more formal
analysis below. However, a large literature has been growing up suggesting that intangible or
institutional factors, not easily quantified, may contribute to the explanation of these differences.
We will attempt later to use some of these together with the more formal analysis, but we can
make a first try here with the raw data and ratios of Tables 1, 2, and 3.
One of the most frequently used measure of these intangibles, published each year, is the
World Competitiveness Index, published each year, that rates countries on a combination of
4
eight factors, labeled as “Openness, Government, Finance, Infrastructure, Technology,
Management, Labor, and Institutions.” (World Economic Forum, 1998). Other analyses have
focussed specifically on the level of corruption in a country, as in the case of Mauro (1995), who
used data collected by a commercial organization, Business International, from ratings by its
analysts and correspondents in the various countries and regions, for use in estimating “country
risk” factors for sale to clients.
If we rank the ten Asian countries by the level of the overall Competitiveness Index, the
average ratio of MOFA gross product to GDP in the five lower ranking countries was 1.5 per
cent and that in the five higher ranking countries was 3.7 per cent. The corresponding averages
for ratios of MOFA to total employment were .11 per cent and 2.25 per cent and for total
affiliate employment, .17 and 2.70 per cent. If we reverse the procedure and rank the countries
by the share of MOFA gross product in GDP, those with high shares had average
Competitiveness indexes of .832 per cent and those with low shares, average indexes of .018 per
cent. If we use the measures of perceived corruption as reported in Mauro, with low corruption
represented by a high rank (close to 1), we find that the highest four countries ranked by affiliate
output shares rank considerably higher (3.5) than the last four (5.6) with respect to corruption as
well. There are only nine countries for this comparison because we do not have ratings for
China. The difference is even sharper for the rankings by growth in affiliate output. The four
countries with the fastest growth averaged 2.75 on the corruption scale while the four with the
slowest growth averaged 6.4 on corruption.
Thus, among the ten Asian countries, low rankings on the Competitiveness Index or high
levels of corruption, or of the perception of corruption, do seem to discourage U.S. FDI, at least
as measured by these simple ratios and ignoring other potential influences that could be
correlated with the competitiveness measures or the corruption indicator. The effect appears to
5
be particularly strong on the level of employment in U.S. affiliates but the discouraging effect
looks to be quite general.
Determinants of Location
The use of ratios to GDP and population to identify particularly attractive or unattractive
locations for investment provides no information as to the reasons for differences among
countries except size. Yet the existence of differences among the ratios implies that there are
other factors at work, and we wish to find out something about what they are.
As pointed out earlier, the number of Asian economies for which data are available is
uncomfortably small for judging which factors are generally influential, and we therefore begin
the analysis by fitting equations to data on all the developing countries for which the necessary
information is available.
It would seem likely that many influences on FDI location are specific to particular
industries. Mineral resources are required for mining industries, labor costs are more important
for labor-intensive than for capital-intensive industries, human capital abundance may be more
important for high-tech than for low-tech industries, and distance to markets more important for
tradable goods industries than for nontradables. We begin here, however, by attempting to
explain aggregate investment, assuming that industry differences will be reflected in different
coefficients for different activity measures, including not only the size measures of Tables 1 and
2, but also specific activities such as exporting and R & D.
The first step in this search for determinants was to fit a set of equations across
developing countries in general for ten aspects of affiliate activity. The independent variables in
these equations are host country market size, as measured by nominal GDP, the growth in host
country real GDP over the previous period (except that nominal GDP was used for the latest
period), real GDP per capita, distance of the host country from the United States, and a measure
of the rate of taxation on U.S. affiliates in the host country. The data for the first three variables
6
were taken from the Summers and Heston Penn World Tables, supplemented by the World
Bank’s World Development Indicators. The distance measure is shipping distance, from the
sources described in Lipsey and Weiss (1974). The measure of tax rates was the sum of income
taxes and other taxes, as a per cent of sales, from the U.S. outward investment surveys.
We expect host country market size and per capita real income to be positively related to
U.S. affiliate size. The effect of per capita income, beyond that of market size, presumably
would reflect an orientation of U.S.-based firms toward goods and services typically purchased
by higher income consumers, or toward intermediate products and capital goods used to produce
such goods and services. The rate of growth in the preceding period should enter with a positive
coefficient if a high rate in the past is a good predictor of high rates in the future. But it might
enter with a negative coefficient if successive rates of growth were negatively correlated, so that
a high rate in one period predicted a low rate in the next period, as if a country tended to revert to
some “normal” long run growth path. Distance from the United States could have both positive
and negative effects on U.S. FDI. A longer distance makes a foreign operation more difficult
and expensive to supervise, and might therefore discourage investment. However, a longer
distance also makes exporting from the United States more expensive, and might therefore make
local production more desirable and encourage investment. Tax rates, if we have measured them
correctly, should presumably have a negative effect on investment.
The coefficients of the ten equations for 1989 and 1994 are shown in Table 4, except
those for the intercepts and the tax variables, omitted to save space. The coefficients for the tax
variables were all far from statistical significance. The insignificance of the tax variables is
surprising, given the strong impact found in other studies of location (for example, Wei, 1997a),
but the difference may reflect the differences in investment measures and country coverage.
Wheeler and Mody (1992) also reported no consistent impact of tax rates on their measure of
investment, U.S. affiliates’ capital expenditures. The fact that the tax rate coefficients in our
7
equations are more frequently positive than negative suggests another possible problem in their
use. We expect them to be negative because we think of them as an exogenous variable to which
U.S. multinationals respond in choosing the location of their operations. An alternative
possibility is that they are themselves endogenous, reflecting the attractiveness of the host
countries, and representing the efforts of the host countries to exact a price from the
multinationals for the use of host country resources. In that case they would not belong in these
equations, and the positive coefficients sometimes found would be there because the tax rates are
acting as proxies for host country advantages not included among the independent variables.
In general, both large market size and high average real income are shown to have
attracted U.S. FDI, whatever the measure of FDI activity used. The only exceptions are MOFA
exports, not positively correlated with GDP, and net exports, negatively, but not significantly,
related to country size. As has been the case in most studies, there is no evidence that low wages,
which would be associated with low per capita real income, were the main magnets for inward
FDI as a whole. Large host country market size was not apparently a magnet for production
destined for markets outside the host country, although it could be if it brought lower costs
gained from economies of scale, as was found for manufacturing production in a much earlier
period in Kravis and Lipsey (1982).
Previous growth in real GDP almost always appears, surprisingly, with a negative sign.
Few of the coefficients are statistically significant at the 5 per cent level in a two-tailed test, but a
couple of them are. Still, the consistency of the negative signs remains puzzling.
Coefficients for distance from the United States are negative. They are not statistically
significant in 1989 but several of them are in the 1994 equations.
We thus have two consistent and statistically significant positive influences on U.S.
affiliate activity, country size, as measured by GDP, and country real GDP per capita. Two
8
influences are consistently negative, but often not statistically significant. One is easily
explainable (distance from the United States) but one contradicts expectations (real past growth)
While the determinants discussed above explain between 40 and 60 per cent of the
variation in most measures of U.S. FDI activity across developing countries, that leaves half of
the variation unexplained. The unexplained variation for the ten Asian countries, as represented
by residuals from the equations for 1994, is shown in Table 6.
The sizes, and even the direction, of the residuals vary among the FDI measures, but
there are some persistent patterns. One is that some countries show positive residuals, indicating
more U.S. FDI activity than is explained by the equations and the variables included, across all
or almost all activities. Others show all or almost all negative residuals, indicating less than the
predicted levels of activity.
Number
Positive
Negative
China
3
7
Hong Kong
1
9
India
0
10
Korea
0
10
Taiwan
1
9
Indonesia
8
2
Malaysia
9
1
Philippines
8
2
Singapore
10
0
Thailand
9
1
Overpredicted
Underpredicted
9
The most striking pattern is that countries with FDI activity higher or lower than
predicted in one respect are correspondingly higher or lower in all or almost all other respects.
Nine out of the ten countries are either high or low in at least eight out of the ten measures of
U.S. FDI activity. The five with higher levels of FDI than predicted by the equations are
Indonesia, Malaysia, the Philippines, Singapore, and Thailand. It is hard to see any obvious
pattern in the list. Those with higher than predicted FDI activity score a little higher on the
overall Competitiveness Index but worse on the perceived corruption scale mentioned earlier,
than those with negative residuals, although the difference is not large. The high FDI group is
not particularly more outward oriented than the others, according to the classification of trade
policies in World Bank (1987).
Some of the patterns of the residuals in Table 6 do suggest explanations. The low
numbers for U.S. FDI in China, India, and Korea probably reflect long periods of hostility
toward direct investment and the restrictions placed on it as a result by host country
governments. The high positive residuals for Singapore and possibly those for Malaysia and
Thailand may reflect active pro-investment government policies and incentives.
There are two plausible interpretations of the residuals from these equations for U.S. FDI
activity. One is that positive residuals represent the effects of some short term influences during
a period that produce more investment than is warranted by the long term situation of a country,
and that negative residuals represent short term influences that, perhaps for political reasons,
such as the U.S. boycott of Viet Nam, or unfamiliarity with a location, hold back potentially
profitable investment. In such cases, the residuals from an equation for one period would enter
equations for the succeeding period with a negative sign. A period of temporarily excessive
investment would be followed by one of lower investment. A level of FDI in a country that was
lower than would be optimal would be followed by a catch-up.
10
Another interpretation could be that a higher or lower than expected level of FDI reflects
not a disequilibrium to be corrected but rather a permanent, or at least long term advantage of a
location that is not incorporated in the explanatory variables. In that case, residuals for one
period would enter equations for the following period with a positive sign.
The two interpretations of the residuals are tested in Table 7. There is no evidence of any
catch-up or correction of previous period disequilibrium. The previous period residuals are
uniformly positive, indicating that a location with unexplained attractiveness or unattractiveness
to U.S. FDI in one period retains that characteristic in the next period. The residuals apparently
represent unincluded characteristics of locations that attract or repel U.S. FDI, but they are
characteristics that we have not so far been able to identify. Another possibility is that they
reflect economies of agglomeration: a large presence of U.S. affiliates in a country paves the way
for other affiliates to enter by accumulating knowledge and lowering costs of entry for followers.
Determinants of Affiliate Characteristics
The same set of independent variables does much less well in explaining the
characteristics of U.S. affiliates in developing countries. While some of the equations and
individual coefficients are statistically significant, the degree of explanation is much lower than
for the size variables and, more important, the explanations are inconsistent from year to year.
The export orientation of affiliates, as measured by the ratio of exports to sales, appears
to be negatively related to the size of the local market and the average host country tax rate.
Both of these relations are in the expected direction, but neither one produces significant
coefficients in all three years. The same relationship holds for the ratio of net exports to sales.
As expected, the average nominal compensation per employee is positively and
significantly related to per capita real income. No equations for capital intensity were
statistically significant and only one out of six for measures of dependence on imports and one
out of three for R & D intensity.
11
The characteristics of U.S. affiliates in the ten Asian countries, described in Table 8,
show some large differences, not necessarily related to the variables in the equations. The very
wide range in the ratio of gross product to the stock of investment, with the highest three times
the lowest, illustrates the problems with using the stock as a proxy for production. The lowest
ratio by far, and therefore the largest exaggeration from using the stock, is for China, although
the reason is not obvious. The other low ratios, for Hong Kong and Singapore, probably reflect
the role of these countries as intermediaries for investment in other countries. The ratios of gross
product to sales show how much of affiliate sales are produced in the affiliate, as compared with
processing of inputs produced elsewhere. The highest ratio by far is for Indonesia, as would be
expected from the concentration in primary production, followed by the Philippines. The lowest
ratios are for Hong Kong and Singapore, with no primary production and extensive entrepot
activities. The exports/sales ratios in 1994 were highest for Indonesia, presumably reflecting the
industrial composition of the investment. The next in line were the two most open economies,
Singapore and Hong Kong, and the two lowest were India and Korea. Capital intensities, as
measured by ratios of fixed capital to employment, were extremely high in Indonesia,
presumably because the petroleum industry is so capital intensive, but the next two countries in
this ranking are Singapore and Hong Kong, perhaps because they are the countries with the
highest incomes, and therefore the highest wage levels. Within assets, Indonesia has the highest
share in fixed assets, while Hong Kong and Singapore, more involved in Finance, have low
ratios, as does Taiwan. The R & D intensities show Taiwan at the top, followed, surprisingly, by
India, a country with extremely little U.S. investment at all, and then by Singapore and Korea.
Again, it seems likely that these ratios cannot be explained without reference to the industry
distribution of the investment.
12
Determinants of Location in Manufacturing
Since the determinants of location should differ among industries, as mentioned earlier, it
is appropriate to move toward disaggregation as far as the data allow. A first step in this
direction is to examine the determinants of manufacturing investment, since that group has some
common features that distinguish it from primary products and possibly also from services.
In Table 9 we show the results of fitting equations for manufacturing in 1994,
corresponding to those for all industries. On the whole, the fit of the equation is a little better,
although not for all aspects of investment, and particularly not for employment and total assets.
As for total investment, the most consistently positive influences on U.S. firms’ activity were
market size and average real income. Coefficients for growth in real income were negative, but
never statistically significant. Those for tax rates were positive, but more often than not,
significant. The major different result was that coefficients for distance from the United States
were generally significant and always negative; the further a country was from the United States,
given its other characteristics, the lower the level of U.S. firms’ manufacturing affiliate activity.
There was no indication that the greater expense of exporting from the United States associated
with greater distance led to higher investment to substitute for exports. The only aspects of
manufacturing affiliate activity that were not influenced significantly and positively by per capita
real income in the host country were employment and net exports. The attraction of a
presumably better educated labor force in higher income host countries was presumably offset to
some extent, in its effect on these variables, by the higher price of labor in those countries.
The residuals from these equations for the ten Asian countries, reported in Table 10,
show some differences from the pattern for total investment. Indonesia, which attracted more
total U.S. investment and activity than predicted by the total FDI equations, attracted less
manufacturing investment than predicted. The difference comes from primary production,
mainly oil. The other difference is that while these Asian countries as a group attracted more
13
total U.S. investment and affiliate activity than predicted by the equations for all developing
countries, they attracted less manufacturing production, although not less manufacturing
employment, than predicted While the residuals among Asian countries for total investment,
shown in Table 6, were almost evenly split between positive and negative (49 positive and 51
negative), those for manufacturing, shown in Table 10, are mostly negative (39 positive and 58
negative). Given market size, real per capita incomes, and distance from the United States, the
Asian countries attracted less U.S. direct investment activity in manufacturing than we would
have predicted. The exceptions were Malaysia, the Philippines, Singapore, and Thailand. No
obvious relation to the institutional variables discussed earlier jumps out from this array.
Tables 11 and 12 report the same information from equations for two major industry
groups within U.S. manufacturing investment, nonelectrical and electrical machinery. The fit is
poorer in these equations and some countries are lost because of lack of information, but a
couple of differences between these industries and manufacturing as a whole stand out. In these
two industries, in which the ratios of exports to sales among U.S. affiliates in developing
countries are much higher than in manufacturing as a whole, market size is of little influence as
an attraction for investment, and distance from the United States is of no consequence. Only per
capita real income is a consistently positive, and usually significant, attraction for U.S. affiliate
activity. Previous period growth in GDP seems to decrease exports, as if affiliates were first
developed for export markets but tended over time to become naturalized, more focussed on host
country markets.
These differences between the two machinery industries and manufacturing as a whole
suggest that there must be some industries with behavior very different from that of the
machinery industries. We examine two of the more important other industries, Foods and
Chemicals, in Tables 13 and 14. In these industries, less geared to export markets, large market
size is a strongly significant attraction for U.S. affiliates, while long distance from the United
14
States is a significant obstacle to investment. The exception to these relationships is net exports
in the Food industry. Large market size was negatively related to net exports and distance from
the United States encouraged net exports. High per capita GDP was generally a positive
influence but usually not a significant one, and high tax rates were positively associated with
most activity measures except exporting. Thus, what attracted U.S. affiliate activity in these
industries was quite different from what attracted those in the machinery industries.
Conclusions and Directions for Research
The most consistent characteristic attracting U.S. direct investment and FDI activity to
developing countries in general was large market size, but even that influence varied among
industries and activities. Large market size did not appear to promote exports by U.S. affiliates
and did not attract U.S. manufacturing investment in the export-oriented nonelectrical and
electrical machinery industries.
High per capita income also attracted U.S. affiliates, and that influence was particularly
notable in the two machinery industries, but was not a major factor in the food and chemical
industries.
The rate of growth of the host economy in the preceding period generally appeared to be
a negative influence on U.S. investment and affiliate activity, a surprising relationship. It was
rarely a statistically significant influence in equations for all industries or for total
manufacturing, but it did appear to have a negative effect on exports in the two machinery
industries and a positive effect on several aspects of affiliate activity in the food industry. A
possible explanation of the negative effect on machinery exports is that rapid growth in the host
country may have caused affiliates to shift their sales from export markets to local host country
markets.
Distance from the United States was another variable with little visible influence on
most affiliate activities, usually negative except on affiliate exports, particularly net exports for
15
all industries as a group. In manufacturing, the effect was stronger and consistently negative.
The negative influence was absent in the machinery industries but strong in foods and
chemicals, except for a positive effect on net exports of chemicals. Thus, there is no evidence
that FDI is placed in distant markets to replace potential exports for which distance imposes
high transport costs. Instead, distance, even crudely measured by shipping distance, seems to
impose costs of operation that discourage investments in those industries specializing in local
sales. However, there was no such effect in industries specialized in exports.
Tax rates on U.S. affiliates do not appear to be a significant influence on FDI in all
industries taken together, but do appear as a paradoxical positive influence in equations for total
manufacturing activity. That relationship too can be traced to the two industries mainly
dependent on local or host country sales, foods and chemicals. The tax rate may be a fee for
access to local markets, a proxy for desirable characteristics of local markets that we have not
identified, but that make them attractive to investors.
Among the ten Asian countries that are the focus of this paper, Singapore and Malaysia
had the highest U.S. affiliate shares of total output, followed by Hong Kong, the Philippines,
and Indonesia. India, China, and Korea were at the low end of the distribution. In terms of
employment shares, Singapore and Hong Kong were at the top, followed by Malaysia, with
India and China again at the bottom. These shares are roughly related to measures of
institutional characteristics, such as those published in World Economic Forum(1998).
The equations across all developing countries do help to explain U.S. affiliate activity in
the ten Asian countries. They also provide a ranking for them by the residuals from the
equations, a measure of the influence of factors not included in the equations, a ranking that is
different from that of the production or employment shares. Given the characteristics described
above and included as independent variables in the equations, Singapore, Malaysia, and
Thailand have the most unexplained U.S. FDI activity; that is, the largest positive residuals
16
(actual minus expected levels), from the equations for many FDI activities. That is the case for
investment stocks, sales, exports, net exports, and R&D. Malaysia is close behind for these
activities and at the top for output, employment, and employee compensation. Korea is the
country with the least FDI activity, considering its other characteristics. That is, Korea had the
largest negative residuals, followed by China, Taiwan, Hong, and India. In general, the
countries with higher FDI activity than expected in one respect are higher in most other respects
as well.
In FDI manufacturing activity by U.S. firms, Singapore is again the most frequent
leader, relative to its attributes, but is surpassed by Malaysia in some respects, and the
Philippines and Thailand also show higher than expected U.S. affiliate activity. All the other
Asian countries show less than the expected activity that would be implied by their
characteristics and the equations for developing countries. China and Korea, and in some
respects, Hong Kong, were among the lowest.
The generality of these measures of attractiveness to U.S. FDI activity is reflected also
in persistence over time. The residuals for one period add substantially to the predictions of
activity for the next period. In other words, while the variables included in the equations
explain about half of the variation in U.S. affiliate across countries, there are other persistent
country characteristics not identified here that explain much of the rest of that variation. A
country with less than expected inward FDI and FDI activity in one period will almost always
be below average ten or fifteen years later
The rankings of Asian countries by their residuals, that is their deviations from the
values implied by the equations, do not appear to have any obvious relation to the institutional
variables mentioned earlier. That fact suggests that the institutional variables themselves, as
measured in these surveys, are partly explained by the same economic factors that explain FDI
activity. That is not to say that no institutional variables are important; it is hard not to suspect
17
that there is some impact of the long period of hostility toward inward FDI in China, India, and
Korea, and the active pro-investment policies of Singapore, Malaysia, and Thailand.
It is clear that differences among industries are an important part of the explanation of
location and characteristics of FDI. A major factor appears to be the distinction between
export-oriented and host country-oriented industries, but other industry characteristics, such as
capital intensity, may be important. The results therefore suggest the desirability of industry
disaggregation in understanding the location decisions of multinational firms and differences in
behavior among affiliates in different countries.
A different strategy would be to go beyond U.S. FDI to examine FDI from other
countries. Since only a few countries report outward FDI data, such a program might involve
shifting from home country data on outward investment to host country data on inward
investment such as that reported in Ramstetter (1996) or the studies of individual studies in
Dobson and Chia (1997). Some comparability across host countries would be sacrificed to
produce more observations and the opportunity to study the combined effect of host-country and
home-country characteristics.
18
References
Dobson, Wendy, and Chia Siow Yue, Editors (1997), Multinationals and East Asian Integration,
Ottawa, Canada, International Development Research Centre, and Singapore, Institute of
Southeast Asian Studies.
Kravis, Irving B. and Robert E. Lipsey (1982), “The Location of Overseas Production and
Production for Export by U.S. Multinational Firms,” Journal of International Economics,
Volume 12, No. ¾, May, pp. 201-223.
Lipsey, Robert E.. and Merle Yahr Weiss (1974), “The Structure of Ocean Transport Charges,”
Explorations in Economic Research, Vol. 1, No. 1, National Bureau of Economic
Research, pp. 162-193.
Mauro, Paolo (1995), “Corruption and Growth,” Quarterly Journal of Economics, Volume CX,
Issue 3, pp. 681- 712.
Ramstetter, Eric (1996), “Trends in Production in Foreign Multinational Firms in Asian
Economies: A Note on an Economic Myth Related to Poor Measurement,” Kansai
University Review of Economics and Business, Vol. 24, Nos. 1-2, March, pp.49-107.
U.S. Department of Commerce (1985), U.S. Direct Investment Abroad: 1982 Benchmark
Survey Data, Bureau of Economic Analysis, Washington, DC, December.
_________________________ (1992), U.S. Direct Investment Abroad: 1989 Benchmark
Survey, Final Results, Bureau of Economic Analysis, Washington, DC, October.
_________________________ (1998), U.S. Direct Investment Abroad: 1994 Benchmark
Survey, Final Results, Bureau of Economic Analysis, Washington, DC, May.
Wei, Shang-Jin (1997a), “How Taxing is Corruption on International Investors?” NBER
Working Paper 6030, Cambridge, MA, National Bureau of Economic Research, May.
19
____________ ((1997b), “Why is Corruption So Much More Taxing Than Tax? Arbitrariness
Kills,” NBER Working Paper 6255, Cambridge, MA, National Bureau of Economic
Research, November.
Wheeler, David, and Ashoka Mody (1992), “International Investment Location Decisions: The
Case of U.S. Firms,” Journal of International Economics, Vol. 33, No. ½, August, pp. 5776.
World Bank (1987), World Development Report, 1987, Johns Hopkins University Press for the
World Bank.
World Economic Forum (1998), Global Competitiveness Report, 1998
20
Table 1
Monetary Measures of the Size of US FDI in Ten Asian Locations, 1994
($ Million)
US MOFAs
Stock
Gross
Product
2,044
9,024
280
6,109
1,727
2,761
1,751
10,722
2,771
2,765
678
4,900
232
4,649
1,452
3,579
1,803
5,750
2,810
2,644
Sales
Assets
Fixed
Assets
Exports
3,225
29,729
983
8,229
5,554
11,579
5,211
46,871
13,690
9,627
5,199
48,237
1,061
13,487
5,098
11,837
4,555
32,164
12,575
10,755
2,093
5,120
317
7,446
1,256
4,484
1,246
5,408
1,783
4,130
705
671
49
2,989
5,217
2,608
12,960
29,063
4,879
1,327
Country
China
Hong Kong
India
Indonesia
Korea a
Malaysia
Philippines
Singapore
Taiwan
Thailand
Non Bank Affiliates of Non Bank Parents
Stock
Sales
Assets
Fixed
Assets
2,455
9,509
470
6,214
2,967
2,867
1,874
10,811
3,314
3,208
4,630
31,015
n.a.
8,871
14,849
12,086
6,622
48,088
15,476
11,348
7,466
51,593
2,271
14,357
15,283
12,681
6,504
33,675
15,460
16,151
2,865
5,796
647
8,029
5,148
4,843
2,392
6,289
3,176
7,140
Country
China
Hong Kong
India
Indonesia
Korea a
Malaysia
Philippines
Singapore
Taiwan
Thailand
a
Republic of Korea
Source: U.S. Department of Commerce (1998).
21
Table 2
Employment Measures of the Size of US FDI in Ten Asian Locations
(000)
All US
Affiliates
Nonbank
Affiliates of
Nonbank Parents
MOFAs
Affiliates
50% or Less
US-owned
China
Hong Kong
India
88
123
51
87
112
48
62
91
18
25
21
30
Indonesia
Korea a
Malaysia
Philippines
Singapore
Taiwan
Thailand
63
65
131
104
106
69
102
61
61
129
94
102
66
100
52
29
121
66
94
59
70
9
32
8
28
8
7
30
Country
a
Republic of Korea
Source: U.S. Department of Commerce (1998).
22
Table 3
Measures of Country Size and Relative US Affiliate Size, 1994
Country
China
Hong Kong
India
Indonesia
Korea, Republic of
Malaysia
Philippines
Singapore
Taiwan
Thailand
GDP
Millions of
US$
540,925
130,801
303,720
176,892
380,820
70,759
64,139
71,039
243,285
143,038
Employment
(000s)
671,990
2,870
360,000
82,039
19,837
7,618
25,166
1,649
8,939
32,095
MOFA Gross
Product as %
of GDP
MOFA Empl.
as % of Total
Employment
Affiliate
Employment
As % of Total
Employment
0.13
3.75
0.08
2.63
0.38
5.06
2.81
8.09
1.16
1.85
0.009
3.171
0.005
0.063
0.146
1.588
0.262
5.700
0.660
0.218
0.013
4.286
0.014
0.077
0.308
1.720
0.413
6.428
0.772
0.318
23
Table 4
Changes in the Size of US FDI in Ten Asian Locations, 1994/1982
US MOFA
Country
China
Hong Kong
India
Indonesia
Korea, Republic of
Malaysia
Philippines
Singapore
Taiwan
Thailand
US FDI
Stock
Gross
Product
4.19
1.87
2.74
10.04
2.35
1.69
7.46
8.74
3.86
5.11
1.01
.74
6.63
2.12
1.68
5.18
4.56
4.02
Sales Empl.
3.96
1.59
.66
9.20
2.68
1.45
3.32
7.33
3.72
2.46
.72
1.00
2.07
2.05
.81
2.19
1.11
3.33
Assets
Fixed
Assets
7.87
2.16
2.06
9.55
3.79
1.72
5.71
9.85
8.40
3.23
2.71
1.84
7.30
2.80
1.74
4.59
4.46
7.17
24
Table 5
Coefficients for Four Explanatory Variables in Equations for
Various Aspects of U.S. Affiliate Activity in Developing Countries
Explanatory Variables
Measures of
U.S. Affiliate Activity
Nominal
GDP
($Million)
2
Growth in
Real GDP
Real GDP
per Capita
R
Distance
from U.S.
(Prob. F)
-.138
(0.7)
-.133
(0.5)
-.192
(0.3)
-.006
(0.9)
-.080
(1.1)
-.136
(0.3)
-.037
(0.2)
.112
(0.4)
.205
(1.0)
-.000
(0.1)
.567
(0.001)
.546
(0.001)
.554
(0.001)
.518
(0.002)
.635
(0.000)
.568
(0.001)
.441
(0.007)
.480
(0.005)
.487
(0.006)
.793
(0.000)
-.495
(2.0)
-.466
(2.2)
-.927
(1.2)
-.016
(2.1)
-.237
(3.0)
-.609
(1.0)
-.157
(0.8)
.162
(0.3)
.622
(1.6)
-.007
(1.5)
.588
(0.000)
.599
(0.000)
.563
(0.000)
.334
(0.016)
.601
(0.000)
.700
(0.000)
.461
(0.002)
.390
(0.014)
.389
(0.014)
.446
(0.008)
1989
Investment stock
Gross Product
Sales
Employment
Employee compensation
Assets
Fixed assets
Exports
Net exports
R & D expenditures
17.2
(5.0)
21.4
(4.8)
39.5
(4.1)
0.61
(5.0)
7.5
(6.0)
28.6
(4.0)
11.3
(3.7)
2.2
(0.5)
-3.5
(1.0)
0.18
(6.5)
-2,298
(1.2)
-2,004
(0.8)
-4,741
(0.9)
-121
(1.8)
-1,021
(1.5)
-2,966
(0.7)
-887
(0.5)
-264
(0.1)
1,433
(0.8)
-29
(2.6)
.367
(3.0)
.339
(2.1)
1.311
(3.9)
.009
(2.2)
.127
(2.9)
1.040
(4.1)
.226
(2.1)
.637
(4.2)
.427
(3.5)
.0020
(2.2)
1994
Investment stock
13.0
(3.9)
Gross Product
12.6
(4.4)
Sales
27.9
(2.6)
Employment
0.39
(3.7)
Employee compensation
5.27
(4.9)
Assets
22.6
(2.7)
Fixed assets
7.4
(2.9)
Exports
0.4
(0.1)
Net exports
-7.3
(1.4)
R & D expenditures
0.21
(3.2)
Note: Figures in parentheses are t-statistics
-1,485
(1.2)
-1,178
(1.0)
-5,296
(1.3)
-10
(0.2)
-320
(0.8)
-7,893
(2.4)
-66
(0.1)
-2,527
(1.0)
-2,011
(1.0)
-50
(1.9)
.454
(4.4)
.325
(3.7)
1.585
(4.9)
.005
(1.5)
.123
(3.7)
1.754
(6.9)
.202
(2.6)
.717
(3.5)
.485
(3.0)
.0074
(3.6)
25
Table 6
Residuals From Equations Across All Developing Countries, 1994
Country
China
Hong Kong
India
Indonesia
Korea, Republic of
Malaysia
Philippines
Singapore
Taiwan
Thailand
Investment Stock
-1,553.82
-921.84
-1,466.33
2,661.18
-4,841.04
2,064.68
618.77
4,440.88
-3,229.56
1,311.82
Country
Assets
China
Hong Kong
India
Indonesia
Korea, Republic of
Malaysia
Philippines
Singapore
Taiwan
Thailand
1,144.37
8,578.57
-3,860.20
5,855.58
-13,671.33
6,555.20
1,560.06
3,403.08
-7,435.25
5,824.98
Sales
-4,406.86
-7,448.83
-2,784.63
2,112.66
-13,734.33
7,180.82
3,584.75
19,656.71
-5,769.56
5,197.12
Fixed Assets
-642.13
-81.80
-2,067.79
3,205.47
-3,142.03
2,835.00
-962.63
1,794.79
-1,950.64
1,596.59
Gross
Product
-2,586.73
-1,481.81
-1,735.25
562.39
-3,873.70
3,437.34
297.35
2,592.21
-1,659.24
1,339.45
Exports
1,558.85
-4,775.68
-801.10
4,209.02
-5,428.59
1,244.78
813.74
14,119.06
-4,580.37
218.87
Employment
-95.70
-31.37
-19.68
10.07
-112.58
133.49
61.78
38.04
-50.02
60.10
Net Exports
4,053.76
-4,546.40
-1,347.42
3,031.07
-3,156.15
-1,405.86
-535.71
12,631.99
-2,850.77
-1,238.31
Employee
Compensation
-1,320.22
-562.45
-238.78
-78.75
-1,334.96
1,237.97
431.23
728.45
-521.73
469.58
R&D
-46.00
-96.47
-20.88
-18.40
-76.05
35.57
15.85
86.32
22.51
6.22
26
Table 7
Coefficients for Explanatory Variables Including Previous Period Residuals
in Equations for Various Aspects of U.S. Affiliate Activity in Developing Countries
Explanatory Variables
Measures of
U.S. Affiliate Activity
Nominal
GDP
($Million)
Growth in
Real GDP
Real GDP
per Capita
Distance
from U.S.
Residual
From
Previous
Period
R
2
(Prob.F)
1989
Investment stock
Gross Product
Sales
Employment
Employee compensation
Assets
Fixed assets
Exports
Net exports
R & D expenditures
22.6
(8.7)
29.3
(10.3)
54.1
(8.0)
0.8
(5.2)
9.2
(14.5)
37.7
(6.7)
15.6
(5.2)
6.3
(1.4)
-1.9
(0.7)
0.2
(19.9)
-1,837
(1.2)
-1,500
(0.9)
-4,091
(1.0)
-110
(1.4)
-740
(2.2)
-2,285
(0.7)
-605
(0.4)
-1,561
(0.6)
483
(0.3)
-17
(5.5)
.288
(3.0)
.212
(2.0)
1.106
(4.4)
.004
(0.8)
.101
(4.3)
.924
(4.4)
.145
(1.3)
.629
(3.6)
.491
(4.9)
.001
(2.5)
-.042
(0.3)
.017
(0.1)
.172
(0.4)
-.002
(0.2)
-.072
(1.8)
.060
(0.2)
.053
(0.3)
.389
(1.4)
.388
(2.4)
-.001
(1.6)
.81
(2.8)
.83
(3.6)
.57
(3.0)
.48
(0.9)
.90
(4.6)
.80
(3.0)
.69
(2.1)
.45
(2.4)
.48
(4.7)
.46
(6.2)
.897
(.000)
.922
(.000)
.891
(.000)
.807
(.001)
.968
(.000)
.872
(.000)
.742
(.003)
.744
(.005)
.887
(.001)
.994
(.000)
-.362
(1.9)
-.352
(2.3)
-.513
(0.8)
-.012
(2.3)
-.212
(2.9)
-.265
(0.6)
-.009
(0.1)
.304
(1.2)
.732
(2.9)
-.004
(1.0)
1.31
(5.1)
.91
(5.5)
1.58
(5.4)
1.26
(6.4)
1.05
(3.8)
1.61
(5.3)
1.12
(5.4)
2.32
(8.6)
2.30
(6.5)
4.72
(4.6)
.829
(.000)
.858
(.000)
.818
(.000)
.781
(.000)
.765
(.000)
.878
(.000)
.796
(.000)
.892
(.000)
.841
(.000)
.792
(.000)
1994
Investment stock
14.6
(6.4)
Gross Product
14.6
(7.9)
Sales
32.9
(4.5)
Employment
.40
(6.3)
Employee compensation
5.8
(6.5)
Assets
26.8
(4.8)
Fixed assets
8.6
(5.3)
Exports
0.9
(0.3)
Net exports
-5.3
(1.7)
R & D expenditures
.4
(5.6)
Note: Figures in parentheses are t-statistics
-1,567
(1.7)
-872
(1.2)
-3,621
(1.2)
-17
(0.7)
-255
(0.7)
-6,980
(3.1)
-350
(0.5)
-1,279
(1.0)
-51
(0.0)
-22
(1.1)
.387
(5.3)
.252
(4.2)
1.292
(5.5)
.003
(1.6)
.104
(3.6)
1.534
(8.4)
.150
(2.9)
.550
(5.5)
.304
(3.0)
.004
(2.0)
27
Table 8
U.S. MOFA Characteristics in Ten Asian Locations, 1994
China
Hong Kong
India
Indonesia
Korea
Malaysia
Philippines
Singapore
Taiwan
Thailand
Gross Prod.
Stock
Gross Prod.
Sales
.332
.543
.829
.761
.841
1.296
1.030
.536
1.014
.956
.210
.165
.236
.565
.261
.309
.346
.123
.205
.275
Exports
Sales
.219
.436
.050
.634
.121
.421
.255
.620
.218
.271
Fixed Assets
Employment
($000)
Fixed Assets
Assets
33.8
56.3
17.6
143.2
43.3
37.1
18.9
57.5
30.2
59.0
.403
.106
.299
.552
.246
.379
.274
.168
.142
.384
R&D
Sales
.22
.17
.51
.06
.31
.23
.27
.36
.80
.03
Source: Table 1 and U.S. Department of Commerce (1998).
28
Table 9
Coefficients for Five Explanatory Variables in Equations for
Various Aspects of U.S. Affiliate Activity in Manufacturing in Developing Countries, 1994
Nominal
GDP
Investment Stock
Gross Product
Sales
Employment
Employee Compensation
Assets
Fixed Assets
Exports
Net Exports
R&D Expenditures
12.2
(5.1)
11.7
(4.7)
27.3
(3.6)
0.34
(3.8)
4.5
(4.8)
21.7
(4.1)
8.7
(5.3)
47.5
(3.5)
51.9
(3.6)
.20
(5.0)
Growth in
Real GDP
-1,034
(1.1)
-1,052
(1.1)
-1,543
(0.5)
-9
(0.3)
-223
(0.6)
-1,769
(0.9)
-517
(0.8)
-4,461
(0.8)
-3,103
(0.7)
-29
(1.5)
Real GDP
Per Capita
.217
(2.9)
.214
(2.8)
.662
(2.8)
.004
(1.3)
.073
(2.5)
.469
(2.9)
.119
(2.3)
1.10
(2.6)
.76
(1.9)
.004
(2.3)
Distance
from U.S.
-.471
(2.6)
-.492
(2.7)
-1.11
(1.9)
-.013
(2.0)
-.202
(2.9)
-.765
(1.9)
-.329
(2.7)
-2.46
(2.4)
-1.78
(2.1)
-.011
(2.9)
Tax
Rate
17.7
(2.3)
22.4
(2.8)
34.9
(1.4)
.226
(0.8)
6.5
(2.2)
33.9
(2.0)
13.1
(2.5)
96.5
(1.9)
57.2
(1.1)
.39
(2.5)
R
2
(Prob.F)
.564
(.000)
.540
(.001)
.419
(.005)
.319
(.020)
.525
(.001)
.472
(.002)
.561
(.000)
.461
(.009)
.553
(.005)
.579
(.001)
Note: Figures in parentheses are t-statistics
29
Table 10
Residuals from Equations for Manufacturing Across All Developing Countries, 1994
($ million except for employment, in thousands)
Investment
Stock
Value
Added
Fixed
Assets
Assets
Sales
Employee
R&D
Employment Compensation Expenditures
Exports
Net Exports
China
-2,878.1
-2,709.5
-1,965.3
-4,561.0
-7,093.7
-82.9
-1,247.1
-62.7
-10,258.9
.
Hong Kong
-2,379.5
-2,325.4
-1,618.2
-6,830.1
-9,117.3
-31.5
-829.6
-13.4
-12,648.4
-9,225.4
-787.0
-710.9
-635.1
-1,588.6
-1,131.5
-17.7
-176.2
-12.6
-1,817.7
-5,540.6
Indonesia
-1,263.5
-1,447.0
-862.9
-1,932.8
-1,500.6
-6.5
-375.4
-18.4
-5,074.5
-4,070.9
Korea, Republic of
-2,330.8
-3,000.3
-1,681.4
-5,784.7
-8,360.1
-96.3
-1,067.6
-54.5
-12,531.6
-14,303.5
Malaysia
2,721.3
2,661.0
2,180.6
5,332.5
7,133.9
127.3
1,114.7
58.9
11,943.9
8,601.8
Philippines
1,076.4
820.8
555.5
1,485.8
3,214.9
55.1
397.3
17.0
3,197.4
3,563.8
Singapore
2,499.2
2,740.8
1,472.1
9,090.9
12,666.6
37.4
772.6
.
16,374.6
15,203.4
Taiwan
-1,013.6
-1,337.2
-731.4
-2,965.5
-3,203.0
-46.7
-349.8
.
-5,255.7
-5,553.1
Thailand
1,357.8
930.9
823.7
1,698.0
2,958.8
54.6
395.7
18.7
4,885.0
2,724.2
India
30
Table 11
Coefficients for 5 Explanatory Variables in Equations for Various
Aspects of U.S. Affiliate Activity in Nonelectrical Machinery in Developing Countries,1994
Nominal
GDP
Investment Stock
Gross Product
Sales
Employment
Employee
Compensation
Assets
Fixed Assets
Exports
Net Exports
R&D
Expenditures
0.653
(2.3)
0.238
(0.4)
-1.609
(-0.4)
0.017
(1.8)
0.235
(1.3)
-0.780
(-0.3)
0.492
(1.8)
-2.628
(-0.5)
-2.821
(-0.6)
0.006
(2.6)
Growth in
Real GDP
-170
(-1.5)
-156
(-0.7)
-652
(-0.4)
-4
(-1.0)
-69
(-1.0)
-415
(-0.4)
-91
(-1.0)
-3,407
(-2.0)
-3,134
(-2.0)
-4
(-3.5)
Real GDP
Per Capita
Distance
from U.S.
0.031
(3.3)
0.038
(2.2)
0.211
(1.7)
0.001
(2.3)
0.014
(2.6)
0.130
(1.8)
0.015
(2.0)
0.738
(2.6)
0.672
(2.6)
0.001
(4.3)
-0.014
(-0.6)
0.018
(0.4)
0.290
(0.9)
0.000
(0.3)
-0.002
(-0.1)
0.160
(0.7)
-0.008
(-0.3)
-0.274
(-0.4)
-0.202
(-0.3)
0.000
(-0.5)
Taxes as a
Ratio of Sales
984
(1.1)
626
(0.3)
-2,309
(-0.2)
15
(0.5)
466
(0.8)
-106
(0.0)
1,038
(1.3)
15,081
(0.7)
14,123
(0.8)
1
(0.2)
R
2
(Prob. F)
0.370
(0.018)
0.126
(0.185)
0.140
(0.207)
0.217
(0.068)
0.213
(0.086)
0.102
(0.262)
0.190
(0.147)
0.495
(0.031)
0.497
(0.030)
0.540
(0.001)
Note: Figures in parentheses are t-statistics
31
Table 12
Coefficients for 5 Explanatory Variables in Equations for Various Aspects of U.S. Affiliate Activity in
Electrical Machinery in Developing Countries, 1994
Nominal
GDP
Investment Stock
Gross Product
Assets
Fixed Assets
Sales
Employment
Employee
Compensation
Exports
Net Exports
R&D
Expenditures
Growth in
Real GDP
Real GDP
Per Capita
Distance
from U.S.
R
2
Taxes as a
Ratio of Sales
(Prob. F)
.488
(0.7)
.598
(1.4)
.279
(0.2)
-.009
(-0.0)
.738
(0.4)
.087
(2.1)
.559
(1.9)
-299
(-1.2)
-142
(-0.9)
-510
(-1.2)
-111
(-0.9)
-560
(-0.8)
1
(0.1)
-43
(-0.4)
0.057
(2.8)
0.032
(2.5)
0.062
(1.8)
0.008
(0.8)
0.106
(1.9)
0.001
(0.5)
0.013
(1.4)
0.048
(1.0)
-0.006
(-0.2)
0.185
(1.9)
0.078
(2.7)
0.185
(1.2)
-0.002
(-0.6)
-0.022
(-0.9)
-9
(-0.0)
-625
(-0.5)
-1297
(-0.4)
-1045
(-1.0)
-3467
(-0.6)
-88
(-0.7)
-296.
(-0.3)
0.3107
(0.0411)
0.2639
(0.0578)
0.2889
(0.0674)
0.3004
(0.061)
0.2542
(0.0103)
0.0482
(0.3163)
0.0998
(0.236)
1.229
(0.7)
.678
(0.6)
.006
(0.8)
-2102
(-3.5)
-1425
(-3.4)
-3
(-1.2)
0.339
(5.2)
0.229
(5.0)
0.000
(1.9)
0.087
(0.5)
0.060
(0.5)
0.000
(0.6)
-1854
(-0.4)
-1326
(-0.4)
15
(0.7)
0.7823
(0.0025)
0.7673
(0.0032)
0.0526
(0.3188)
Note: Figures in parentheses are t-statistics
32
Table 13
Coefficients for 5 Explanatory Variables in Equations for
Various Aspects of U.S. Affiliate Activity in the Food Industry
2
Nominal
GDP
Gross Product
Assets
Fixed Assets
Sales
Employment
Employee
Compensation
Exports
Net Exports
R&D
Expenditures
1.888
(5.5)
3.697
(4.9)
1.026
(5.2)
4.701
(4.5)
0.042
(4.0)
0.528
(4.3)
0.249
(1.2)
0.185
(0.9)
0.011
(4.9)
Growth in
Real GDP
35
(0.2)
246
(0.8)
191
(2.5)
554
(1.4)
3
(0.7)
78
(1.4)
229
(2.9)
235
(2.9)
2
(2.1)
Real GDP
Per Capita
0.031
(2.2)
0.048
(1.6)
0.010
(1.7)
0.044
(1.4)
0.000
(0.9)
0.006
(1.2)
-0.003
(-0.5)
-0.004
(-0.6)
0.000
(1.2)
Distance
from U.S.
Taxes as a
Ratio of Sales
-0.091
(-3.4)
-0.227
(-3.9)
-0.069
(-4.1)
-0.328
(-3.7)
-0.003
(-3.3)
-0.035
(-3.8)
0.002
(0.2)
0.009
(0.6)
-0.001
(-3.8)
2,980
(2.6)
6,786
(2.6)
2,124
(3.2)
9,099
(2.6)
67
(2.0)
943
(2.3)
467
(0.7)
476
(0.7)
22
(2.9)
R
(Prob. F)
0.6606
(0.0001)
0.6427
(0.0004)
0.7223
(0.0001)
0.6193
(0.0007)
0.4487
(0.0028)
0.6003
(0.0004)
0.2846
(0.0311)
0.2863
(0.0305)
0.6249
(0.0001)
Note: Figures in parentheses are t-statistics
33
Table 14
Coefficients for 5 Explanatory Variables in Equations for
Various Aspects of U.S. Affiliate Activity in the Chemical Industry
2
Gross Product
Assets
Fixed Assets
Sales
Employment
Employee
Compensation
Exports
Net Exports
R&D
Expenditures
Nominal
GDP
Growth in
Real GDP
Real GDP
Per Capita
3.422
(5.8)
4.713
(5.5)
1.598
(5.1)
6.088
(4.8)
0.051
(7.1)
1.006
(5.6)
0.264
(1.7)
-0.331
(-2.7)
0.095
(4.0)
-172
(-0.8)
-295
(-0.9)
-162
(-1.3)
-173
(-0.4)
-3
(-0.9)
-18
(-0.3)
-13
(-0.2)
-10
(-0.2)
-1
(-0.1)
0.033
(1.8)
0.063
(2.4)
0.023
(2.4)
0.066
(1.7)
0.000
(2.0)
0.010
(1.8)
0.005
(0.9)
-0.001
(-0.3)
0.001
(1.2)
Distance
from U.S.
Taxes as a
Ratio of Sales
-0.109
(-2.5)
-0.183
(-2.9)
-0.060
(-2.6)
-0.259
(-2.8)
-0.002
(-3.3)
-0.046
(-3.4)
-0.004
(-0.3)
0.020
(2.2)
-0.006
(-3.0)
4,360
(2.3)
10,717
(3.6)
3,624
(3.4)
12,522
(2.8)
66
(2.9)
1,661
(2.9)
30
(0.1)
-630
(-1.6)
86
(1.1)
R
(Prob. F)
0.5898
(0.0002)
0.6640
(0.0001)
0.6257
(0.0002)
0.5809
(0.0006)
0.6923
(0.0001)
0.5980
(0.0001)
-0.0228
(0.5089)
0.2254
(0.0621)
0.4019
(0.0094)
Note: Figures in parentheses are t-statistics
34